phunk

Dear OC' Guru's
I'm currently working on a Gigabyte EP35-DS3 mainbord with a corsair dual channel kit CM2X1024-8500C5D (it's on the Gigabyte QVL list)
In single channel they work fine, no errors with memtest86+ v2.11
No errors after several passes.
When placed in Dual channel all hell breaks loose, random reboots, lots of errors within seconds with memtest.
I've set the timing and voltage correctly, but still same problem.
(The settings as adviced by "the ram guy" from the corair company)
Cas settings Cas 5-TRCD 5-TRP 5- TRAS 15
ACT to ACT Delay(tRRD) 5
Rank Write To READ Delay 5
Write To Precharge Delay 15
Refresh to ACT Delay 68 [Auto]
Read to Precharge Delay [Auto]
Static tRead Value [Auto]
Static tRead Phase [Auto]

However when I downgrade my memory to 800 everything works fine.

Even corsair doesn't have a clue what's causing the problems, the already replaced the RAM with brand new one's. So for now their solution is to replace them again. But i'm not that sure that will help...

So as far as I know it must be a DRAM timing setting.
I asked Gigabyte for some help, but I don't think even they understand the advanced DRAM timing settings, Cause my tech support question remains unanswered.
Hope you guys can help me out.
Many thanks already
Grtz

leeghoofd

Are you running any OC ? Could be a mobo divider related issue... Strap selected ? MCH voltage ? better take some piccies of ya bios screens and link us to them plz...

__________________

Gamer

Up the DDR2 voltage a bit, could help.

__________________

phunk

I added some 'piccie's' from the BIOS setup screen.
The settings in the pics are until now, the only one working in Dual channel memory config, unfortuantly @ 800 instead of 1066
I'm running currently no OC

Would it be safe to increase the voltage ? by, say 0.1v

Attached Thumbnails

wutske

Upping the voltage with 0.1V is safe. In your screenshot your memory voltage is already running 0.3V higher than normal.

//edit: the increased voltage (+0.3V) is normal for running the memory at 1066MHz, these modules are designed to run at 2.1V instead of 1.8V.

leeghoofd

I would try to set Static tRead value 7-8, MCH OverVoltage Control +0.1, FSB OverVoltage +0.1 Probably the bios reads out the SPD wrongly and sets too tight timings...

If you set 1066Mhz and try with one dimm ? does that work ? If so swap the dimm for the other one and see if it works... if both pass try another dimm slot... and so on

__________________

phunk

I tought it might damage the hardware, cause corsair's advise is to run them at the voltage they are designed for, but I'll give it a go.

Then everything works fine, I explored every possibility.
even 2 dimm's in single channel are working fine.
Dual channel was tested in slot 1 & 3 and in slot 2 & 4 -->same problem

I'll give it a go, and will start a memtest when I'll go of to work.

Many thanks for all the advice so for.
I should have been a member for years on this forum

leeghoofd

I fit works in single channel it's more of a NB problem... maybe the little bit more MCH voltage will do the trick. Could be it runs too hot too (place a small fan over it)

Too bad I don't see a strap setting in them screens, could be this is just a faulty divider... Have you tried flashing another bios ? Or tried another version of MemTest ?

__________________

thorgal

I never really tested much DDR2 on P35/P45, but on P965 there were some issues with the first 5 subtimings (ACT to ACT -> Read to precharge).

Try putting them all except "Refresh to act" to 10 cycles instead if 4-4-8-4. Not saying this is definately going to work, just mentioning this because in some cases it did work

__________________

phunk

I've tried everything you guys suggested, but the problem persists.
-Latest BIOS version
-other version memtest
-increased the voltage of DRAM, CMH, FSB
-Modules in other slots
-set the first 5 subtimings to 10 (exept refresh to act delay)
-set Static tRead value to 7-8

I think we have to conclude it's a crappy mainboard ??